Real-Time Precision Needle Tracking for Gynecological Interstitial Brachytherapy: A Phantom Feasibility Study

Abstract

Listen

Translate article

To determine if an advanced needle navigation and real-time precision tracking system developed for prostate biopsies can be used for the purpose of real-time guidance of interstitial needles in modern image-guided gynecologic brachytherapy. An agarose gel pelvic phantom was previously designed using 3D Slicer and Autodesk Fusion360 to model an average uterus from a 50-patient dataset, a vaginal canal from KleenSpec (TM, Welch Allyn) speculum dimensions, and a rectum to accommodate a BK 8848 endocavity biplane ultrasound [US] probe. After manufacturing the phantom with a 7.2cm x 3.7cm x 4.4cm (LxWxH) uterus and an attached 2.7cm x 1.4cmx 1.7cm (LxWxH) focal HR-CTV target, the model was imaged using MRI. The uterus and HR-CTV were contoured on MRI to aid in registration with US and provide a target for needle placement. The tracking system with endorectal probe was then used to image the phantom. Ultrasound images from a 180-degree sweep were acquired and a 3D Ultrasound-based volume was generated. Using the commercially available system, the MRI was imported and an MRI-Ultrasound fusion performed based on the uterus contours and then refined using contour-based deformable registration. Three desired locations for needle placement were defined to test the ability of the system to guide needles at these locations: one in the most lateral aspect of the HRCTV, one at the HRCTV-uterus interface, and one inside the uterus. The needle tracking sleeve that was indexed to the probe was used in conjunction with the software interface to guide the 17-gauge titanium needle placement in real time. The distance between actual needle and the intended needle location was measured for each successful needle placement. When possible, needles were removed and reinserted at the same desired location to test the reproducibility of insertion under US guidance. Two of the three intended needles were successfully placed using real-time US guidance. The most lateral needle in the HRCTV was not inserted due to limitations of the phantom design, specifically inaccessibility to the agarose phantom through the phantom wall supports. The inserted needles were visible and tracked to the desired location within 0.6 mm (needle located at HRCTV-uterus interface), and 2.2 mm (needles inserted centrally in the uterus). For the target within the uterus, needles were reinserted at the same target of interest with a ∆ 0.15mm in greatest range between attempts. We have demonstrated a proof of concept in which existing technology for prostate biopsies can be utilized for the application of interstitial gynecologic brachytherapy with a high degree of fidelity. Ongoing modifications to our validation phantom and the tracking system are needed to allow for variations in uterine anatomy, more rigorous evaluation of reproducibility, and multi-catheter placement.